The potential of remote sensing in ecological status assessment of coloured lakes using aquatic plants Sebastian Birk a, b, *, Frauke Ecke b, c a Faculty of Biology, Department of Aquatic Ecology, University of Duisburg-Essen, Universitätsstraße 5, 45141 Essen, Germany b Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences, Box 7050, SE-75007 Uppsala, Sweden, c Department of Fish, Wildlife and Environmental Studies, Swedish University of Agricultural Sciences, Skogsmarksgränd, SE-90183 Umeå, Sweden A R T I C L E I N F O Article history: Received 18 January 2014 Received in revised form 18 June 2014 Accepted 26 June 2014 Keywords: Biological metric Boreal zone Eutrophication Macrophytes Surveillance monitoring Unmanned aircraft system A B S T R A C T Field-based survey methods for aquatic vegetation have been identified as resource-demanding. Recent advances in remote sensing (RS) with sub-decimetre resolution allow for surveying aquatic vegetation at the species level. Coloured lakes, mainly due to high concentrations of humic substances, are globally widespread. However, high colour impedes the identification of submerged vegetation via remote sensing. Here, we evaluate the potential of using only emergent, floating and floating-leaved taxa that are detectable by high-resolution RS (RS-taxa) to assess the ecological status of lakes. In a dataset covering 72 Swedish low alkaline coloured lakes, we identified 31 RS-taxa. The power of RS-taxa assemblages to predict non-RS assemblages was analysed by a combination of ordination (Detrended Correspondence Analysis, DCA, and Redundancy Analysis) and multiple regression analysis. We compared the performance of a trophic metric score based on RS-taxa with that based on field data from all taxa along different environmental gradients. Forty percent of the variability of the first non-RS taxa DCA-axis was predicted by the DCA-results based on RS-taxa. Correlations of the trophic metric score and total nitrogen concentrations were equally strong for the dataset based on RS-taxa compared to the dataset based on all taxa. For total phosphorous concentrations, the correlation was stronger for the dataset based on all taxa, but for a complex water quality gradient (including sulphate, N-species, chlorophyll and percent cover of wetlands in the riparian buffer) the correlation was higher for the RS-taxa dataset. The significant linkage between the two community fractions (remotely-sensible and non-sensible) revealed considerable assemblage concordance, suggesting a notable potential of the use of remote sensing in lake macrophyte monitoring. The established trophic metric score seems most qualified for surveillance monitoring that, in combination with the eased efforts of data acquisition, detects long-term changes of the aquatic environment caused by shifts in climate, land use and (related) eutrophication. ã 2014 Elsevier Ltd. All rights reserved. 1. Introduction Aquatic vegetation has a central role in the structuring and functioning of aquatic systems by, for example, providing food and habitat, retaining and cycling of nutrients and affecting sediment redox conditions (Jaynes and Carpenter 1986; Engelhardt and Ritchie 2001; Wetzel 2001). Due to their known response to environmental conditions, sampling and monitoring of macro- phytes is an integrative part of ecological assessments in different parts of the world including Europe (European Union, 2000), U.S.A. (EPA, 1998) and Australia (reviewed by Hart et al., 1993). Transect and belt transect methods are among the most frequently used sampling methods for macrophytes in lakes (EPA, 1998; Natur- vårdsverket, 2010a; Kanninen et al., 2013). Using these methods, the time effort to sample macrophytes in lakes <5 km 2 sums up to approximately four days (two days for two people) (Naturvårds- verket, 2010a). Considering the c. 100,000 lakes >0.04 km 2 in Scandinavia (Henriksen et al., 1998), assessment of ecological status based on field methods is a major logistic and financial challenge. Remote sensing allows for the acquisition of synoptic data uniformly covering large areas in space and time, repeatedly and nonintrusively (Green et al., 1996), making its use in environmental monitoring highly convenient (Barrett and Curtis,1992; Lunetta and Elvidge, 1999; Anker et al., 2013). Remote sensing by manned aircrafts equipped with RGB (red green blue) and IR (infrared) sensors has been used for mapping aquatic vegetation since the * Corresponding author. Tel.: +49 201 183 3113; fax: +49 201 183 4442. E-mail addresses: sebastian.birk@uni-due.de (S. Birk), Frauke.Ecke@slu.se (F. Ecke). http://dx.doi.org/10.1016/j.ecolind.2014.06.035 1470-160X/ ã 2014 Elsevier Ltd. All rights reserved. Ecological Indicators 46 (2014) 398–406 Contents lists available at ScienceDirect Ecological Indicators journa l home page : www.e lsevier.com/loca te/ecolind